Baldur Bjarnason reposted
“In the report, the auditor general points out multiple concerning examples of mistakes in those summaries that could have a direct and negative impact on a patient’s subsequent care.”
Who's Who of AI
Baldur Bjarnason
Writer, web developer and consultant based in Hveragerði, Iceland. Lapsed Interactive Media Academic. Webby Tech Stuff and webby book stuff.
https://www.baldurbjarnason.com/
https://softwarecrisis.dev/
What they're sharing
Your doctor’s AI notetaker may be making things up, Ontario audit finds - Ars Technica arstechnica.com
View on Bluesky →
programmablemutter.com
hollywoodreporter.com
View on Bluesky →
spectrum.ieee.org
scientificamerican.com
2ndbreakfast.audreywatters.com
nooneshappy.com
pivot-to-ai.com
sonjadrimmer.com
pivot-to-ai.com
pivot-to-ai.com
dotart.blog
Articles & links
“AI Isn't Management. Try Explaining That to Matthew Prince” https://www.programmablemutter.com/p/ai-isnt-management-try-explaining > At last we have created the Corporation That Eliminates Middle Managers from classic management text Don't Eliminate the Middle Managers!
Baldur Bjarnason reposted
if you post anything on youtube, you can and should opt your channel out of this www.hollywoodreporter.com/business/dig...
“Hidden Signals Can Hijack AI Voice Systems - IEEE Spectrum” https://spectrum.ieee.org/voice-ai-audio-attacks > Research shows sounds unheard by human ears can hijack models’ behavior The entirety of the “AI” ecosystem has been built in perfect ignorance of basic security prac…
AI Weekly's analysis
→
- Adversarial audio attacks operate above human hearing range, making them undetectable to users and standard monitoring tools.
- The attack surface expands as voice AI shifts from answering queries to executing autonomous actions in enterprise and infrastructure contexts.
- Existing audio security tooling cannot detect ultrasonic adversarial command injection in current voice AI pipelines.
Read full analysis →
“AI Is entering health care, and nurses are being asked to trust it | Scientific American” https://www.scientificamerican.com/article/ai-is-entering-health-care-and-nurses-are-being-asked-to-trust-it/
“Of Course They Booed” https://2ndbreakfast.audreywatters.com/of-c/ > It's this sneering attitude, I'd argue, that is driving so much of the pushback against "AI" and against ed-tech
“The AI Bubble — No One's Happy” https://nooneshappy.com/article/the-ai-bubble/ > The buildout is underwritten by an implicit guarantee that none of the people who made the bet will be the ones who pay for it.
“The SpaceX IPO works like a crypto fraud, but with AI – Pivot to AI” https://pivot-to-ai.com/2026/05/28/the-spacex-ipo-works-like-a-crypto-fraud-but-with-ai/ > SpaceX is doing a crypto-style fraud, but on the real stock market.
“"Employers Want Applicants with AI Skills": How Hearsay Manufactures Reality — Sonja Drimmer” https://sonjadrimmer.com/blog-1/2026/5/22/employers-want-applicants-with-ai-skills-how-hearsay-manufactures-reality > Of all the ads for positions I clicked, none said a single thing…
“OpenAI user numbers go flat — just in time for the IPO – Pivot to AI” https://pivot-to-ai.com/2026/05/22/openai-user-numbers-go-flat-just-in-time-for-the-ipo/ > But “operating income margin” was not great
“McClatchy runs AI slop with journalists’ names on it – Pivot to AI” https://pivot-to-ai.com/2026/05/27/mcclatchy-runs-ai-slop-with-journalists-names-on-it/
“The Stepford AI — cobbles” https://dotart.blog/cobbles/the-stepford-ai > The LLMs aren't real. But the misogynistic desires and harms to all of us are real.
Their own posts
Recent commentary
Finding productive uses for LLMs while at the same time doing nothing to mitigate the numerous harms makes things WORSE not better. Responding to “AI” critique by saying that they have their uses and with work we can make them genuinely productive is not a rational response. It’s accelerationism
One of the annoying consequences of LLMs is that many people now seem eager to dismiss any and all writing that people enjoy as “obviously LLM-generated”, just because it uses em-dashes or writing structures with decades if not centuries of history in English.
We’re at the stage of the bubble where the elderly and retired start to talk about the amazing revolution that is “AI” and start buying “AI training courses for the elderly” because they think they’re missing out on a tech purpose-designed to make office and creative industry workers unemployed
So tempted to find one of those “signs of LLM writing” lists and just make sure that everything I write intentionally checks all the boxes just to fuck with people.
Somebody’s running an ad campaign in Icelandic radio promoting free courses on how to use “AI”. The target audience seems to be retirees and older people.